Method for producing steel in a metallurgical vessel
By employing a calculation model to estimate exhaust air system pressure using process data and visual inputs, the method addresses the maintenance and accuracy issues of traditional pressure sensors, enhancing gas flow control and recovery in steel production.
Patent Information
- Application Number
- EP2023185250
- Authority / Receiving Office
- EP · EP
- Patent Type
- Patents
- Current Assignee / Owner
- Priority Date
- 2022-08-02
- Filing Date
- 2023-07-13
- Publication Date
- 2024-11-13
- Estimated Expiration
- 2043-07-13
AI Technical Summary
The existing methods for controlling gas flow in steel production, particularly in metallurgical vessels, rely on pressure sensors in harsh environments, leading to complex and costly maintenance due to dust contamination and inaccurate readings.
A calculation model is used to determine exhaust air system pressure based on process data, including oxygen flow, carbon monoxide and dioxide concentrations, temperature, additives, and visual data from cameras, eliminating the need for direct pressure measurement.
This approach simplifies the gas flow system, reduces maintenance costs, and optimizes exhaust gas recovery by accurately controlling gas intensity, maximizing carbon monoxide yield while minimizing energy loss.
Smart Images

Figure IMGF0001 
Figure IMGF0002
Abstract
Description
[0001] The invention relates to a method for producing steel in a metallurgical vessel, in which crude steel is melted in the metallurgical vessel and oxygen is introduced into the crude steel, wherein an exhaust air system is arranged above the metallurgical vessel in which a predetermined target pressure is maintained in a regulated manner.
[0002] In steel production, for example in a converter (especially in a BOF - Basic Oxygen Furnace), the removal of exhaust gases through the exhaust air system is generally controlled by means of a pressure measurement in the exhaust gas duct.
[0003] US 2017 / 0335417 A1 proposes setting the pressure setpoint using a value that takes into account the size of the flame spot, which is captured by a camera aimed at the converter opening. Additional information, such as the position of the skirt and the amount of oxygen injected into the converter, is also used to set the setpoint. However, the actual pressure is determined by measuring it using a pressure sensor located in the exhaust air system.
[0004] WO 2020 / 212782 A1 expands the use of cameras for electric arc converters (EAF converters) and proposes using not only the size but also the intensity of the flame spot. Here, too, the actual pressure is measured using a pressure sensor located in the exhaust air system.
[0005] The current state of the art therefore controls the gas flow through the exhaust air system by measuring the pressure in the exhaust air system (cooling stack). Due to the harsh environment, the pressure gauge required for this is complex and requires a relatively large installation space. Furthermore, due to the dusty gas, the pressure measurement sockets are prone to contamination and blockage, leading to inaccurate pressure measurements. Consequently, the maintenance costs for the system are also relatively high.
[0006] The invention is based on the Task The aim is to further develop a method of the type mentioned above in such a way that pressure measurement in the exhaust air system becomes unnecessary, thereby eliminating the problem described above.
[0007] The solution to this problem by the invention is characterized in that the actual pressure present in the exhaust air system is determined by means of a computer model to which current process data are supplied, wherein the computer model determines the actual pressure present in the exhaust air system from stored data of previously performed processes and for this purpose establishes a correlation between the process data and the actual pressure in the exhaust air system, wherein the process data at least comprise: the volume flow of oxygen supplied into the metallurgical vessel and the concentration of carbon monoxide and / or carbon dioxide in the exhaust air system.
[0008] The process data preferably also includes the temperature at a defined point of the exhaust system, the type and amount of additives added to the metallurgical vessel, the size of the gap formed between the top of the metallurgical vessel and the bottom of the exhaust system, and the degree of wear of the metallurgical vessel.
[0009] It is particularly preferred that the process data further comprise at least one analyzed image of a camera which records a flame or smoke which is generated during the process, in particular between the top side of the metallurgical vessel and the bottom side of the exhaust air system (i.e. in the region of the above-mentioned gap).
[0010] Accordingly, no pressure measurement is performed in the exhaust air system using a pressure sensor. Therefore, a pressure sensor, as is required for this purpose in the state of the art, is not required.
[0011] The calculation model is preferably based on a logistic regression between the process data and the actual pressure in the exhaust air system. In statistics, logistic regression refers to regression analyses for modeling the distribution of dependent discrete variables. In this case, a statistical correlation is established between the determined process parameters and the resulting actual pressure in the exhaust air system. Accordingly, it is not necessary to measure the pressure in the exhaust air system itself; rather, this is determined indirectly from the other process parameters by the calculation model.
[0012] The data stored in the computational model from previously performed processes are preferably supplemented and / or modified with currently acquired process data. This approach is known as "machine learning" in the field of artificial intelligence and therefore needs no further description here.
[0013] The metallurgical vessel is in particular a converter (BOF - Basic Oxygen Furnace) or an electric arc furnace (EAF).
[0014] According to the invention, the previously common pressure measurement in the exhaust air system is replaced by a calculation model based on available process data; in addition, other process parameters can also be taken into account.
[0015] The inventive concept thus proposes to replace the pressure in the exhaust air system (exhaust duct) directly measured in the previously known solutions with a value calculated from the available data previously recorded during the steel production process as well as currently determined (measured) parameters. the amount of oxygen injected into the converter, the concentration of carbon monoxide (CO) and carbon dioxide (CO 2 ) in the gas mixture leaving the converter, the temperature of the cooling stack (or at a defined point in the exhaust air system), the amount and type of admixed materials, the position of the skirt and the degree of wear of the metallurgical vessel
[0016] Any combination of the above-mentioned process variables or parameters can be considered.
[0017] Additional information for the pressure calculation is the size of the flame spot captured by the camera looking at the converter orifice, as well as the amount of smoke escaping from the vessel, which is also determined from the camera images.
[0018] The proposed method thus provides optimal control of the process gas flow during steelmaking, especially during the oxygen blowing phase, which is preferably used in the blast furnace (BOF).
[0019] This involves the use of a computational model (i.e., a computer running a corresponding algorithm) that utilizes a statistical model. This model can be adapted to historical data (i.e., data obtained in the past during processes performed at this metallurgical plant), particularly using the well-known method of machine learning.
[0020] In particular, when obtaining said data for the calculation model, the results of previous synchronous measurements of the pressure with simultaneous recording of the above-mentioned parameters are taken into account, ie a collection of previous data is carried out which allows statements to be made about a correlation of the prevailing actual pressure in the exhaust air system for given values for the above-mentioned parameters.
[0021] The parameters of the computational model can be adjusted by iterative optimization of an objective function that describes the deviation of the model predictions from the actual pressure measurements.
[0022] The proposed method makes it possible to reduce the complexity of the gas flow system and its maintenance costs. The decision to increase or decrease the gas flow rate (by increasing or decreasing the fan speed accordingly) is essentially based on knowledge of the pressure in the exhaust duct. Since pressure sensors in the exhaust air system are deliberately omitted due to the harsh environment, the invention provides for determining the actual pressure in the exhaust air system without directly measuring it, but rather using the aforementioned calculation model.
[0023] An advantage is that the proposed approach also increases flue gas recovery. The steel production process in a blast furnace plant is accompanied by the release of large quantities of smoke, carbon dioxide (CO2), carbon monoxide (CO), and other gas components. Insufficient flue gas intensity leads to an increase in the amount of smoke and CO. Excessive flue gas intensity, on the other hand, leads to the combustion of CO in the flue gas duct and thus to a reduction in the energy value of the gas mixture. The method proposed according to the invention optimizes the flue gas intensity and maximizes the CO yield.
[0024] The drawing shows an embodiment of the invention. Fig. 1 schematically shows a metallurgical vessel with an exhaust air system in which pig iron is produced, and Fig. 2 schematically shows the determination of the current pressure in the exhaust air system using a computer model to which current process data are fed.
[0025] In Figure 1 a metallurgical vessel in the form of a BOF is shown, above which an exhaust air system (exhaust gas system) 2 is arranged in a known manner.
[0026] Steel production from pig iron in vessel 1 occurs by oxidizing excess carbon and other impurities with oxygen O 2 , which is blown through the molten pig iron. The oxygen blowing lance 11 required for this purpose is immersed in the interior of vessel 1. This process releases large quantities of smoke, carbon dioxide (CO 2 ), carbon monoxide (CO), and other gases.
[0027] The exhaust air system 2, consisting of an apron 12 and a gas duct connected to it, is installed above the vessel 1 to clean and collect the flue gases. The cleaned, stored gas mixture is later used for energy recovery by converting CO into CO2. A gap 4 is formed between the top of the vessel 1 and the bottom of the exhaust air system 2, specifically the apron 12. The size of the gap is available as a measured value SP.
[0028] The concentration of carbon monoxide CO or carbon dioxide CO 2 can be measured via the measuring system 9. Similarly, the volume flow rate at which the oxygen O 2 is introduced into the crude steel via the blowing lance 11 can be determined via a measuring system 10.
[0029] As mentioned above, inadequate primary gas capture leads to an increase in smoke and gas emissions from the BOF area and increases the risk of toxic gases such as CO. On the other hand, additional intake of ambient air into the primary gas stream leads to the combustion of CO in the flue system, thus reducing the energy value of the accumulated gas mixture, increasing the heat load on the cooling stack, and resulting in an additional gas flow rate, which increases the required intake capacity.
[0030] Devices such as fans 14 and / or flue gas cleaning devices 13, which are installed in the exhaust air system, regulate the intensity of the exhaust gas. According to the state of the art, the decision made by the pressure control device 15 to increase or decrease the gas flow rate is based on measuring the actual pressure p actual in the exhaust air system using a pressure sensor.
[0031] This pressure is compared to the target pressure p setpoint, which ensures a high CO concentration in the accumulated gas mixture and an acceptable amount of dust and sludge emitted. If the measured pressure is higher than the target pressure, the fan power is increased, and vice versa.
[0032] The disadvantage of pressure measurement as provided in the prior art is – as explained – a complex and space-consuming design, which is necessary due to the harsh environment. Pressure measurement becomes inaccurate, and the sensors must be cleaned regularly, leading to high maintenance costs.
[0033] This is where the present invention comes in, providing a new method for controlling the gas flow that replaces direct pressure measurement with a calculation based on parameters of the steelmaking process and, if necessary, visual data obtained with a camera 5 that observes the flame at the container neck. To capture optimal images of the flame or smoke, illumination with a lamp or laser 8 can be used. An image processing unit 6 evaluates the image captured by the camera 5.
[0034] Instead of directly measuring the actual pressure p actual in the exhaust air system 2, the proposed concept relies on the use of a calculation model 3, which is fed with current process data (which, however, does not include the current actual pressure in the exhaust air system). From this current data, as well as on the basis of previously determined and stored data, the actual pressure is determined and output by the calculation model 3. In the exemplary embodiment, it is provided that a pressure calculation unit 7, in which the calculation model 3 is integrated, is fed with image data BD, the size of the gap SP (distance between the container neck and the skirt), the concentration of carbon monoxide CO in the exhaust air system, the concentration of carbon dioxide CO 2 in the exhaust air system, and the volume flow of supplied oxygen O 2 (see Figure 2 ) in order to determine the actual pressure p lst and output it to the pressure control device 15.
[0035] In the example, in addition to the parameters of the amount of oxygen injected and the concentration of carbon monoxide and carbon dioxide, additional parameters of the steelmaking process are also taken into account, which are used to calculate the actual pressure in the exhaust air system. Furthermore, the amount of added substances and the degree of wear of vessel 1 must also be considered.
[0036] For this purpose, the image processing unit 6 analyzes the images taken by the camera 5 of the gap 4 between the converter mouth and the exhaust air system (cooling shaft or apron of the vessel) to determine the size of the bright area generated by the flames. The images are recorded with a CMOS and / or infrared camera. With the help of the light source 8 (lamp; laser), the amount of smoke escaping from the vessel is determined from the images by absorption and / or reflection. All of the above-mentioned process parameters, which are regularly measured during steel production, can be fed into the pressure calculation unit 7, which estimates the pressure p actual in the exhaust line. The pressure calculated by the computational model is used to adjust the exhaust gas intensity instead of direct pressure measurement and is fed to the pressure control device 15.
[0037] Insufficient exhaust intensity leads to an increase in the amount of smoke escaping between the outlet of the BOF converter and the inlet of the exhaust system (cooling stack). The amount of smoke not captured by the primary gas extraction system is estimated based on the intensity of the image from the installed light source (lamp, laser) or the intensity of the light emitted by the light source and reflected by objects in the vicinity of the converter. The value representing this intensity is used to calculate the pressure in addition to the process parameters listed above.
[0038] A statistical model is used as the basis for calculating the pressure depending on the otherwise recorded process parameters, with particular emphasis on logistic regression (see the above explanations).
[0039] Within the framework of the calculation model, pressure is calculated as follows: Each of the above-mentioned recorded process parameters is first normalized: Its mean value (i.e., the mean of the data acquired and stored so far) is subtracted from its (current) value, and the result is divided by its standard deviation. The mean and standard deviation of the parameter are thus determined from historical data (i.e., data previously recorded during the process).
[0040] The normalized parameters are then multiplied with the model parameters and added. Figure 2 where this is illustrated schematically.
[0041] How to continue in Figure 2 As can be seen, the result of the summation is given to the input of the logistic function, ie to the calculation model 3.
[0042] The starting value of the logistic function is the normalized actual pressure. To obtain the absolute actual pressure, the normalized actual pressure is multiplied by the standard deviation of the pressure and added to the mean pressure value. The mean and standard deviation of the pressure are again determined from historical data. List of reference symbols:
[0043] 1Metallurgical vessel (BOF, EAF) 2Exhaust system (exhaust gas duct) 3Calculation model 4Gap between the top of the metallurgical vessel and the bottom of the exhaust system 5Camera 6Image processing unit 7Pressure calculation unit 8Lamp / laser 9Measuring system for the concentration of CO / CO2 10Measuring system for the volume flow of oxygen 11Oxygen blowing lance 12Apron 13Flue gas cleaning device 14Fan 15Pressure control device p Soll specified target pressure in the exhaust air system p Slst actual pressure in the exhaust air system BDImage data SPSize of the gap 4 COCarbon monoxide CO 2 Carbon dioxide O 2 Oxygen
Claims
1. Method for producing steel in a metallurgical vessel (1), in which crude steel is melted in the metallurgical vessel (1) and oxygen is introduced into the crude steel, wherein an exhaust air system (2) in which a predetermined target pressure (pSoll) is maintained by regulation is arranged above the metallurgical vessel (1), characterised in that the actual pressure (plst) present in the exhaust air system (2) is determined by means of a computer model (3) to which instantaneous process data are fed, wherein the computer model (3) determines the prevailing actual pressure (plst) in the exhaust air system (2) from stored data of methods performed earlier and with respect thereto produces a correlation between the process data and the actual pressure (plst) in the exhaust air system (2), wherein the process data comprise at least: - oxygen fed to the volume flow in the metallurgical vessel (1) and - the concentration of carbon monoxide (CO) and / or carbon dioxide (CO2) in the exhaust air system (2).
2. Method according to claim 1, characterised in that the process data further comprise the temperature at a defined point in the exhaust air system (2).
3. Method according to claim 1 or 2, characterised in that the process data further comprise the kind and quantity of additives fed to the metallurgical vessel (1).
4. Method according to any one of claims 1 to 3, characterised in that the process data further comprise the size of the gap (4) formed between the upper side of the metallurgical vessel (1) and the lower side of the exhaust air system (2).
5. Method according to any one of claims 1 to 4, characterised in that the process data further comprise the degree of wear of the metallurgical vessel (1).
6. Method according to any one of claims 1 to 5, characterised in that the process data further comprise at least one analysed image of a camera (5) recording a flame or the smoke occurring during the process, particularly between the upper side of the metallurgical vessel (1) and the lower side of the exhaust air system (2).
7. Method according to any one of claims 1 to 6, characterised in that there is no measurement of the pressure in the exhaust air system by means of a pressure sensor.
8. Method according to any one of claims 1 to 7, characterised in that the computer model (3) is based on a logistical regression between the process data and the actual pressure (plst) in the exhaust air system (2).
9. Method according to any one of claims 1 to 8, characterised in that data, which are stored in the computer model (3), of methods performed earlier are enhanced and / or modified by instantaneously determined process data.
10. Method according to any one of claims 1 to 9, characterised in that it is performed in a metallurgical vessel (1) in the form of a converter (BOF - basic oxygen furnace) or in the form of an electric arc furnace (EAF).
Citation Information
Patent Citations
Method and apparatus for controlling the pressure in the exhaust gas duct of a converter
US20170335417A1
Method to control exhaust fumes aspiration during a steelmaking process
WO2020212782A1
